Adenovirus is a large DNA virus whose natural host is human cells. Virtually every adult has been infected with adenovirus at some time, the major effect being cold-like symptoms. Adenovirus is referred to as a "DNA tumor virus" purely because of its oncogenic effect in rodents. Expression of the adenovirus genome occurs in two stages. First, the early gene products are expressed which encode the E1A and E1B genes. These products are necessary for expression of the late gene products. Late gene products encode proteins necessary for replication as well as the viral structural proteins.
The proteins encoded by the E1A gene of adenovirus have been studied primarily from two points of view. First, the 243 amino acid and 289 amino acid forms of E1A (arising from alternative splicing of the precursor RNA such that the 243 amino acid protein is a subset of the 289 amino acid protein) are both transcriptional regulatory proteins, J. Flint, T. Shenk, Ann. Rev. Gen. 23:141-161 (1989). Secondly, these proteins facilitate the oncogenic transformation of certain rodent cells by other oncogenes, H. E. Ruley, Nature 304:602-606 (1983), and, as such E1A is generally classified as an oncogene. The interactions of E1A proteins with various defined cellular proteins are believed to mediate its oncogenic effects in rodent cells, J. M. Howe, et al., Proc. Natl. Acad. Sci. 87:5883-5887 (1990).
In human cells, E1A does not appear to cooperate with other oncogenes to transform cell lines, however, mild tumorigenic effects have been reported in only two instances. For example, embryonic kidney, F. Graham, J. Smiley, W. Russell, R. Nairn, J. Gen. Virol. 36:59-72 (1977), and embryonic retinoblast cell lines, P. J. Byrd, K. W. Brown, P. H. Gallimore, Nature 298:67-71 (1982), stably expressing E1A are weakly tumorigenic when injected into mice. Moreover, these transformants arise very rarely after cotransfection of E1A with another oncogene.
The normal and transformed phenotypes of a cell are thought to be maintained by a balance of activities encoded by oncogenes and antioncogenes (tumor suppressor genes). The absence of functional tumor suppressor gene products contributes to carcinogenesis. Conversely, replacement or overexpression of these products promises to be useful in controlling tumor cell growth. Only a limited number of tumor suppressor genes have been characterized at present, notably, retinoblastoma protein, Huang et al., Science 24:1563-1566 (1988), and p53, Chen et al., Science 250:1566-1570 (1990). However, when reintroduced into tumor cells, both of these gene products simply kill the cells.
Two reports in the literature claim tumor suppression effects of E1A, but both of these effects arose from epiphenomena not related to the growth controlling effects of E1A expression. Rat sarcoma cells stably transfected with E1A are less tumorigenic in nude rats than their untransfected counterparts (T. A. Walker, B. A. Wilson, A. M. Lewis, J. L. Cook, Proc. Natl. Acad. Sci., 88:6491-6495 (1991). Decreased tumorigenicity of E1A-expressing cells resulted from sensitization of these cells to cytolysis by natural killer cells and these cells were still tumorigenic in nude mice. The second report (D. Yu, K. Scorsone, M. C. Huang Molec. Cell Biol. 11:1745-1750 (1991) showed that E1A prevented an oncogene known as neu from transforming the mouse cell line 3T3. E1A accomplished this solely by inhibiting the expression of neu at the transcriptional level. 3T3 cells transfected with a chimeric neu gene that utilized an E1A-resistant gene promoter were still transformed by neu even in the presence of E1A. The use of E1A to suppress tumor proliferation in the manner implied by the observations above would require E1A expression in all of the cells in the tumor. This requirement poses extreme practical limitations on the use of tumor suppressor genes for therapy since the technology does not exist to introduce a desired gene into 100% of the cell population. In light of these drawbacks, strategies for inhibiting the uncontrolled replication of hyperproliferative cells using tumor suppressor genes have not yet emerged.
Thus, there exists a need to regulate hyperproliferative cell populations through small subpopulations of cells endowed with the ability to dominantly influence their growth. The present invention satisfies this need and provides related advantages as well.